|
on Contract Theory and Applications |
| By: | Dongsoo Shin (Santa Clara University); Roland Strausz (HU Berlin) |
| Abstract: | We study an agency model with vertical hierarchy---the principal, the prime-agent and the sub-agent. The principal faces a project that needs both agents' services. Due to costly communication, the principal receives a report only from the prime-agent, who receives a report from the sub-agent. The principal can directly incentivize each agent by setting individual transfers (insourcing), or sets only one overall transfer to an independent organization in which the prime-agent hires the sub-agent (outsourcing). We show that insourcing is always optimal when the principal can perfectly process the prime-agent's report. When the principal's information process is limited, however, outsourcing can be the prevailing mode of operation. In addition, insourcing under limited information process is prone to collusion between the agents, whereas no possibility of collusion arises with outsourcing. |
| Keywords: | information process; sourcing policy; vertical structure; |
| JEL: | D86 L23 L25 |
| Date: | 2024–02–14 |
| URL: | http://d.repec.org/n?u=RePEc:rco:dpaper:495&r=cta |
| By: | Tao Lin; Yiling Chen |
| Abstract: | We study a repeated Bayesian persuasion problem (and more generally, any generalized principal-agent problem with complete information) where the principal does not have commitment power and the agent uses algorithms to learn to respond to the principal's signals. We reduce this problem to a one-shot generalized principal-agent problem with an approximately-best-responding agent. This reduction allows us to show that: if the agent uses contextual no-regret learning algorithms, then the principal can guarantee a utility that is arbitrarily close to the principal's optimal utility in the classic non-learning model with commitment; if the agent uses contextual no-swap-regret learning algorithms, then the principal cannot obtain any utility significantly more than the optimal utility in the non-learning model with commitment. The difference between the principal's obtainable utility in the learning model and the non-learning model is bounded by the agent's regret (swap-regret). If the agent uses mean-based learning algorithms (which can be no-regret but not no-swap-regret), then the principal can do significantly better than the non-learning model. These conclusions hold not only for Bayesian persuasion, but also for any generalized principal-agent problem with complete information, including Stackelberg games and contract design. |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2402.09721&r=cta |
| By: | Masaki Aoyagi; Maxime Menuet |
| Abstract: | This paper studies the optimal disclosure of information about an agent’s talent when it consists of two components. The agent observes the first component of his talent as his private type, and reports it to a principal to perform a task which reveals the second component of his talent. Based on the report and performance, the principal discloses information to a firm who pays the agent the wage equal to his expected talent. We study incentive compatible disclosure rules that minimize the mismatch between the agent’s true talent and his wage. The optimal rule entails full disclosure when the agent’s talent is a supermodular function of the two components, but entails partial pooling when it is submodular. Under a mild degree of submodularity, we show that the optimal disclosure rule is obtained as a solution to a linear programming problem, and identify the number of messages required under the optimal rule. We relate it to the agent’s incentive compatibility conditions, and show that each pooling message has binary support. |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:dpr:wpaper:1232&r=cta |
| By: | Pablo A. Celhay; Paul Gertler; Marcelo Olivares; Raimundo Undurraga |
| Abstract: | We examine the effect of performance monitoring in public procurement through the lens of organizational culture in a principal-agent model where the manager (principal) and buyers (agents) may have different beliefs about how much the government values efficiency. We show that the effect of performance information not only increases efficiency but is greater when the buyer’s belief is stronger than the manager’s belief. We leverage a new e-procurement system in Chile to test these ideas by randomizing monthly reports on the purchasing performance of buyers and further whether the individual performance reports were disclosed to managers. We find that the reports generated sizable reductions in overspending — with savings reaching a 15% reduction or 0.1% of GDP — but only when individual performance was observable to managers. This is consistent with extrinsic motivation rather than intrinsic motivation driving buyer behavior. Consistent with the theoretical model, we also find that the gain in efficiency is concentrated in procurement units where buyer belief that the government cares about efficiency is stronger than manager belief. Our results highlight the key role played by organizational culture in mediating the impact of purchasing performance information on preventing the misuse of public resources. |
| JEL: | D23 |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:32141&r=cta |
| By: | Kristian Blickle; Zhiguo He; Jing Huang; Cecilia Parlatore |
| Abstract: | We study specialized lending in a credit market competition model with private information. Two banks, equipped with similar data processing systems, possess “general” signals regarding the borrower's quality. However, the specialized bank gains an additional advantage through further interactions with the borrower, allowing it to access “specialized” signals. In equilibrium, both lenders use general signals to screen loan applications, and the specialized lender prices the loan based on its specialized signal conditional on making a loan. This private-information-based pricing helps deliver the empirical regularity that loans made by specialized lenders have lower rates (i.e., lower winning bids) and better ex-post performance (i.e., lower non-performing loans). We show the robustness of our equilibrium characterization under a generalized information structure, endogenize the specialized lending through information acquisition, and discuss its various economic implications. |
| JEL: | D43 D44 D82 G21 G23 L10 |
| Date: | 2024–02 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:32155&r=cta |